The summer of 2026 has turned hydrology from an operating variable into a regional energy-security crisis. Serbia’s Đerdap 1, the largest hydropower plant in the country, was producing only about 5,000 MWh per day in late July—roughly one-third of its normal output—as prolonged heat and drought pushed the Danube toward historic lows.
May and June were reportedly the plant’s weakest production months since it opened in 1970. By early August, inflows were expected to approach 1,500 cubic metres per second, near the biological minimum. Low water also restricted barge loading to 30–40% of capacity and complicated cooling at Serbia’s Kostolac coal complex. One climatic event was therefore simultaneously reducing renewable generation, fuel transport and thermal-plant resilience.
The episode exposes a weakness in how SEE has traditionally valued hydropower. Large plants have been treated as dependable domestic resources that reduce fuel imports and provide flexible generation. They still perform those roles, but historical output is becoming a less reliable guide to future production. Hotter summers, lower river flows and more variable precipitation can reduce energy volumes precisely when cooling demand is high and solar output falls away in the evening.
This does not make hydropower less valuable. It makes flexibility, reservoir management and climate adaptation more valuable. Reservoir plants can still respond quickly to changes in demand, while run-of-river facilities remain exposed to whatever water is available. Operators will need better seasonal forecasting, coordinated management across borders and operating rules that recognise ecological flows as binding system constraints rather than optional environmental considerations.
The Danube is a shared energy artery. Reduced output at Đerdap affects Serbia’s import needs and therefore prices and flows in Romania, Bulgaria, Hungary and the Western Balkans. At the same time, low river levels threatened cooling-water supplies at Romania’s Cernavodă nuclear plant and reduced output at Hungary’s Paks facility. The 2026 crisis demonstrates that a diversified generation mix can still contain correlated climate risks when several technologies depend on the same river basin.
Investment priorities should change accordingly. Refurbishing turbines can extract more electricity from each unit of water, while digital controls can improve dispatch. Batteries can conserve hydro reservoirs for longer shortages by managing intra-day volatility. Additional interconnection can spread local shocks across a wider market. None of these measures creates water, but each can reduce the economic impact of scarcity.
New hydropower proposals also need stress testing against future hydrology rather than twentieth-century averages. Forecast generation, debt service and environmental impacts can all change if drought years occur more frequently. A project that appears inexpensive per megawatt may prove costly per delivered megawatt-hour if water availability is overstated.
SEE’s 2026 lesson is therefore not that hydropower has failed. It is that water can no longer be treated as a constant. The region’s hydro fleet remains one of its most valuable sources of renewable flexibility, but its future contribution will depend on climate-resilient operation, modernised equipment and stronger integration with storage and cross-border markets.
